By way of definition, the heart is said to be in normal sinus rhythm (NSR) when the atria and ventricles beat in synchrony at a heart rate lower than a defined tachycardia heart rate that provides sufficient cardiac output of oxygenated blood to the body at rest and during exercise or stress. The term bradycardia refers to an abnormal slow rate of one or more heart chamber that inappropriately provides insufficient cardiac output at rest or during stress or exercise. The term “tachyarrhythmia” refers to any abnormal fast rhythm of one or more heart chamber that reduces cardiac output and may be amenable of conversion to NSR by “cardioversion” or “defibrillation” or the application of certain anti-tachycardia pacing therapies to the heart chamber as described further herein. Atrial tachyarrhythmias include atrial tachycardia (AT) and atrial flutter or fibrillation (AF) originating from one or more ectopic sites in the right or left atria. Ventricular tachyarrhythmias include ventricular tachycardia (VT) and ventricular flutter or fibrillation (VF) originating from one or more ectopic sites in the ventricles. Supraventricular tachycardia (SVT) can also result from high rate atrial tachyarrhythmias or junctional depolarizations conducted to the ventricles including AV re-entrant tachycardia, which usually conducts down the AV node and up through left postero-lateral bypass tract is considered an SVT. Individuals whose hearts go into VF or into high rate, polymorphic VT can suffer sudden cardiac death (SCD) unless the rhythm terminates either spontaneously or therapeutically within a very short time after onset of such high rate VT or VF.
AF and VF are characterized by chaotic electrical activity that exhibits highly variable depolarization wavefronts that are propagated in directions that differ from the directions of propagation during NSR and more rhythmic tachycardias. The depolarization waves traversing the atria during AF and the ventricles during VF do not follow normal conduction pathways and can vary in direction from beat to beat. During AF and VF episodes (particularly at onset and during the initial phase before cardiac activity diminishes), the depolarization waveforms are irregular in amplitude and hence in appearance when viewed on an electrocardiogram strip or display and are characterized as “polymorphic”. In addition, the atrial or ventricular EGM does not exhibit a characteristic baseline of little electrical activity separating P-waves or QRS complexes, respectively.
The QRS complexes of rhythmic atrial and ventricular tachycardia episodes typically exhibit a regular or “monomorphic” P-waves or QRS waveforms that simply become narrower as heart rate increases from NSR and that are separated by a baseline interval. However, the QRS complexes during certain VT episodes can be polymorphic, particularly from one beat to the next. Such polymorphic VT episodes may be due to reentry conduction through diseased tissue, which results in QRS depolarization wavefronts that are also typically propagated in directions that differ from those prevalent during NSR or monomorphic VT or SVT episodes.
In the field of automatic implantable arrhythmia control devices, particularly ICDs (also referred to as pacemaker/cardioverter/defibrillators or PCDs), the terms “cardioversion” and “cardioverter” and “defibrillation” and “defibrillator” generally refer to the processes of and devices for discharging relatively high energy electrical shocks into or across cardiac tissue to arrest a life threatening tachyarrhythmia. In practice, the conversion of AT or VT or low rate AF or VF to normal sinus rhythm by a relatively low amplitude cardioversion shock delivered in timed synchrony with a sensed atrial or ventricular cardiac depolarization (P-wave or R-wave) is typically referred to as “cardioversion”. The conversion of malignant AF or VF by the same or higher energy shock delivered without requiring such synchronization is typically referred to as “defibrillation”. Synchronization can be attempted, but therapy is delivered without synchronization if synchronization is not possible in a short time. Cardioversion shocks that may or may not be synchronized with a cardiac depolarization or rhythm and may be applied to arrest a VT with a lower range energy pulse of around 1-15 Joules or VF with a medium to high energy pulse of 7-40 Joules, nominally. In the following description and claims, it is to be assumed that the terms cardioversion and defibrillation are interchangeable, and that use of one term is inclusive of the other, unless specific distinctions are drawn between them in the context of the use. For convenience, cardioversion and/or defibrillation shocks or shock therapies are referred to herein as C/D shocks or shock therapies.
Bradycardia cardiac pacing functions are also currently incorporated into ICDs to supplant some or all of an abnormal heart's natural pacing function by delivering appropriately timed pacing pulses to cause a chamber or chambers of the heart to contract or “beat”, i.e., to “capture” the heart. Either single chamber (atrial or ventricular) pacing functions or dual chamber (atrial and ventricular) pacing pulses are applied to the atria and/or the ventricles in response to bradycardia or dissociation of the atrial and ventricular heart rates at a pacing rate to restore cardiac output that is appropriate to the physiologic requirements of the patient. Most recently, synchronized right and left heart pacing, particularly synchronized pacing of the right and left ventricles, has been incorporated into ICDs for heart failure patients who are also susceptible to tachyarrhythymias.
In addition, anti-tachycardia pacing capabilities have been incorporated into ICDs for delivering bursts of pacing pulses or single overdrive pacing pulses to the atria and/or the ventricles to counter and convert certain slow AT or VT episodes to normal sinus rates. The number, frequency, pulse amplitude and width of burst pacing pulse therapies can be programmed by remote programming and telemetry equipment to meet physiologic needs of the particular patient and power conservation requirements.
Among the most important functions of such ICDs are to detect tachyarrhythmias, to correctly identify the tachyarrhythmia, to supply an appropriate cardioversion/defibrillation or burst pacing therapy, and to determine whether or not the supplied therapy was effective.
The typical VT and VF detection criteria that have been employed in commercially released ICDs employ rate/interval based timing criterion and duration or frequency criterion as a basic mechanism for detecting the presence of and distinguishing between tachyarrhythmias. To this end, the intrinsic heart rate is measured on a beat-to-beat basis by timing the R-R interval between successive ventricular sense (VSENSE) event signals output by an R-wave sense amplifier. The measured R-R intervals are compared to a fibrillation detection interval (FDI), a fast tachycardia detection interval (FTDI) and a slow tachycardia detection interval (TDI), and respective VF, fast VT or slow VT counts are accumulated depending on the result of the comparison. One of VF, fast VT or slow VT is declared when a count matches a particular number of intervals required for detection (referred to herein as “NID”). Each rate zone may have its own defined NID, for example, “VFNID” for fibrillation detection, “FVTNID” for fast VT detection, and “VTNID” for slow VT detection.
For example, the measured R-R intervals are compared to the FDI criterion, and the ventricular sensed event is declared a VF event or a non-VF event depending upon the results of the comparison. VF is provisionally declared when the count meets (i.e., equals or exceeds) the VFNID frequency criterion. Similarly, the ventricular sensed event can be declared a fast VT or a slow VT depending on the results of the comparison to the FTDI and the TDI.
Often, SVT episodes causing the ventricles to beat at a rate that meets the FDI and can be inappropriately detected as VF episodes. In ICDs having dual chamber, atrial and ventricular, sensing capabilities, further strategies have been generally followed to detect and classify atrial and ventricular tachyarrhythmias. Algorithms have been developed that identify atrial sensed events from P-waves and/or ventricular sensed events from R-waves and derive atrial and/or ventricular event intervals and/or rates therefrom. Various detection and classification systems have been proposed as set forth in commonly assigned U.S. Pat. Nos. 5,342,402, 5,545,186, 5,782,876, and 5,814,079, that invoke a hierarchy of prioritization rules to make a decision as to when a cardioversion/defibrillation therapy is to be delivered or withheld. These rule-based methods and apparatus have been incorporated in dual chamber ICDs to distinguish atrial and ventricular tachyarrhythmias employing “PR logic” in dual chamber MEDTRONIC® GEM® DR ICDs.
Single chamber ICDs for distinguishing VF from VT or SVT and providing ventricular C/D shock therapies and/or burst pacing therapies do not have the capabilities of sensing P-waves to detect atrial sensed events and analyzing the relationship between atrial sensed events and ventricular sensed events based on detected R-waves. Therefore, many other proposals have been made to examine electrogram (EGM) waveform characteristics, particularly unique waveform characteristics of the QRS complex during NSR, VT, VF and SVT.
One method of discriminating between VF and NSR EGM waveforms as set forth in commonly assigned U.S. Pat. No. 5,312,441, for example, is based on measurements and comparisons of the width of the QRS complex to VF width criterion. A normal QRS complex is generally narrower than the abnormal QRS complex during VF, and so QRS width can be employed to distinguish the normal QRS complex from the abnormal QRS complex during VF. However, there are cases when an abnormal QRS complex during VT will have a different morphology than the normal QRS complex while remaining narrow. Conversely, the QRS complex during certain SVT episodes can also be wide. In those cases, a more sensitive and selective method is needed to discriminate between different waveforms.
As noted above, QRS depolarization waves traversing the ventricles during VF do not follow normal conduction pathways and can vary in direction from beat to beat, whereas QRS depolarization waves during SVT that follow normal conduction pathways or during VT emanating from stable ectopic depolarization sites do not vary significantly in direction of propagation. Therefore, various proposals have been made to distinguish VF from a stable VT or SVT as a function of the QRS wave propagation direction on a beat-to-beat basis.
The VT/VF discriminator disclosed in commonly assigned U.S. Pat. No. 5,193,535 employs two sense electrode pairs, e.g., a near field or bipolar electrode pair and a far field or unipolar electrode pair, that are coupled to detection circuitry for identifying the points in time when the sensed electrical signals resulting from the passage of a depolarization wavefront (QRS complex) meet certain predetermined criteria, hereafter referred to as the first and second “fiducial points”, that may or may not be the same. The cumulative variability of the time intervals separating the occurrence of the first and second fiducial points over a series of R-R intervals that satisfy VF or VT detection criteria is determined. In general terms, the cumulative variability of a series of true VF QRS complexes resulting in satisfaction of VF detection criteria is higher than the cumulative variability of a series of stable VT QRS complexes or SVT QRS complexes satisfying the VF detection criteria. The cumulative variability value or index is used to distinguish VF from high rate VT to trigger or withhold delivery of a C/D shock therapy. Similar techniques are disclosed in U.S. Pat. No. 5,810,739.
A further approach to the discrimination of normal heart beats from abnormal heart beats employing the morphology of the QRS complex is based on making a comparison of the waveform of the QRS complex during tachyarrhythmia with the waveform of a “template” recording of a QRS complex in NSR and optionally, other template recordings made during VF or VT. An ICD is disclosed in commonly assigned U.S. Pat. No. 5,447,519 that discriminates between monomorphic ventricular tachyarrhythmias, particularly VT, from polymorphic ventricular tachyarrhythmias, particularly VF. A fiducial point of each successive QRS complex is detected (e.g., a VSENSE) prompting the storage of sampled and digitized waveform amplitude data within a timing window bridging the point in time of fiducial point detection. Stored sets of such sampled wave shape data are compared data point to data point resulting in a sampled morphology index value for each compared set. The magnitude of the sampled morphology index value or a series such index values are analyzed to determine the presence of a single or a progression of beat-to-beat waveform changes indicative of a polymorphic single transition or progression of QRS complexes from monomorphic to polymorphic waveforms indicative of an arrhythmia that should be treated with aggressive C/D shock therapies. The ICD is preferably provided with a closely spaced and widely spaced pairs of electrodes for sensing each QRS complex as in the above-referenced '535 patent. The closely spaced electrode pair is coupled to sense detect circuitry for identifying the fiducial point and to counting and comparison circuitry for developing rate and onset data. The widely spaced pair of electrodes is coupled to sense and digitizing circuitry for developing the sampled waveform amplitude data from which the morphology index values are derived.
The common approach for such morphology analysis is Correlation Waveform Analysis (CWA) or its less computationally costly counterpart, so-called Area of Difference (AD) analysis. Both require minimization of a function describing difference between two signals (sum of squared differences of wave data points for the case of CWA, and the sum of absolute values of the differences for AD analysis). However, such computations, as typically performed, are more computationally costly and consume more power to carry out than is generally desirable within ICDs.
As set forth in the U.S. Pat. No. 5,439,483, a great deal of information for characterizing the EGM signal, particularly the QRS complexes, if the information can be extracted and analyzed employing mathematical transforms. The Fourier transform is most commonly employed in waveform analysis to find the probability of any individual frequency occurring in the waveform. In this way, a time varying signal is represented as the sum of its frequencies. A large amount of data from a signal may be compressed, and certain information that may be hidden in the data may be viewed from a different perspective. The power of this representation diminishes when the signal that one is trying to represent changes its character unpredictably during the course of the signal. Essentially, local information is lost when the global representation of a Fourier transform is attempted.
As asserted in the '483 patent, an improved method of performing this type of transform is known as the windowed Fourier transform, wherein the time series is divided into small windows in time or in space depending on the nature of the data. The transform is performed to obtain the Fourier spectra of the data at various windows. The problem with this technique is that the uncertainty principle begins to set in. The smaller the window, so designed to better handle the localization of the data, the worse its frequency information becomes. The uncertainty principle can be minimized using the Gabor transform, which makes use of the theorem that the minimum uncertainty is achieved with the Gaussian window. Thus, a Gaussian function is used instead of performing the transform to break the signal down into its basic frequencies represented by a time series of sine and cosine functions. This improves the Fourier transform but still cannot give the detailed information of the time series.
Uses of the Haar wavelet transform for performing morphologic analysis and discrimination of normal and abnormal QRS complexes is described in the above-referenced '483 patent and in commonly assigned U.S. Pat. No. 6,393,316. Wavelets represent a signal in a way that provides local frequency information for each position within the signal or digitized sample of the signal amplitude, as described in detail in the '316 patent. Thus, the wavelet transform can be used to extract information of the time series that is not restricted to the sine or cosine functions of the Fourier transform. Essentially, any function can be chosen that is appropriate for obtaining the relevant information of the time series. The advantage is that the signal can be observed at any time scale, i.e., the technique can zoom in on the signal, up to its finest resolution. As set forth in the above-referenced '316 patent, the wavelet transform is a representation of a signal as a sum of so-called wavelets or little waves. The wavelets are highly localized in time or, in the mathematical language, have compact support. The main difference between the wavelet functions used in wavelet transforms and the sine and cosine functions used in the Fourier transform is that families of wavelets have limited time support that scales exponentially.
There are certain computational advantages of using wavelet transforms instead of Fourier transforms as described in detail in the above-referenced '316 patent. The wavelet transform applied to digitized QRS amplitude sample values will yield a set of wavelet transform coefficient (WTC) data, and a selected sub-set of the WTC data can be employed to accurately represent the QRS complex, and thus will achieve a high degree of information compression. This can be especially important for IMDs because the information compression provided can be employed to substantially reduce the number of required computations. The QRS signal can also be efficiently filtered and de-noised by keeping a number of large amplitude wavelet coefficients and deleting lower amplitude wavelet coefficients. Thus, the use of a wavelet transform-based morphology analysis method significantly reduces the amount of computation necessary to perform the task and performs de-noising of the signal at no additional computational cost.
The above-referenced '483 patent discloses a system and method for characterizing the “seriousness” of a VF episode using wavelet transforms of each QRS complex in a series of QRS complexes, so that a C/D shock may be delivered to the ventricles with an appropriate amount of energy. A wavelet transform of the second derivative of the time series of a fibrillation event is performed on digitized QRS complexes in the EGM to provide spectral functions of the QRS complexes. In the algorithm disclosed in the '483 patent, the results are analyzed for “missing peaks” in the data. The more serious the VF, the more peaks will be missing from the data. This information may then used to modify the energy of the C/D shock as a function of the estimated seriousness of the VF.
The above-referenced '316 patent discloses a method and apparatus for reliable discrimination between ventricular depolarizations resulting from normal and abnormal propagation of depolarization wavefronts through the chambers of a patient's heart by means of a Haar wavelet transform-based method of analysis of QRS complexes of the EGM. Several embodiments are described in the '316 patent that involve the development of WTC templates of NSR as well as SVT QRS complexes and comparison of current high rate QRS complexes satisfying VT or VF rate criteria to the stored WTC templates. A first disclosed embodiment compares template and unknown waveforms in the wavelet domain by ordering WTC data of the template and unknown waveforms by absolute amplitude and comparing the resulting orders of the WTC data. The second and third disclosed embodiments perform analogs of CWA and AD computations in the wavelet domain. All three methods produce good discrimination of QRS complexes during VT episodes from normal QRS complexes during SVT episodes and may be readily implemented in the embedded environments of ICDs. It is asserted that the disclosed embodiments may also be usefully applied to discriminate between other cardiac waveforms in the EGM, including normal atrial P-waves and those associated with atrial AF and AT episodes. Certain features of the wavelet morphology algorithms disclosed in the '316 patent are employed in the single chamber MEDTRONIC® Marquis® VR ICDs.
Both the complexity and the indications for implantation of the above-described ICDs have increased remarkably over the years. Patients who receive such ICDs are typically identified as survivors of SCD secondary to VF that may originate as VT. In such cases, the cost and complexity of such ICDs is deemed warranted. However, many patients likely to suffer SCD are presently un-diagnosed and do not survive their first VF episode. It is believed that certain patient populations exist that could be identified from other indicia and could benefit from a “prophylactic”, low cost, limited function, ICD that simply provides protection against SCD due to VF. To minimize cost of the ICD and the implantation procedure, such a prophylactic ICD would necessarily have limited functions and the capability of delivering only a limited number of high-energy C/D shocks in response to a detected VF episode.
In a prophylactic ICD application, there is concern that the selected patients will exhibit nearly the same frequency of SVT episodes but far fewer polymorphic VT or VF episodes than is exhibited by the conventional ICD patient population. Therefore, it is feared that the use of the current VF detection algorithms will result in a higher percentage of inappropriate VF shock therapies than in the conventional ICD population. This is expected because Bayes' theorem teaches that detection performance depends not only on the detection algorithm's intrinsic performance, but also depends on the population of tachyarrhythmias that the algorithm processes.
In the prophylactic ICD application, AF episodes that conduct rapidly to the ventricle (rapidly conducted AF) are of particular concern. The ventricular rate of such AF events is often similar to that of VF and very hard to discriminate from simultaneous AF and VT/VF on the basis of intervals alone. Wilkoff et al. identified rapidly conducted AF as one of the primary algorithmic causes for inappropriate VT/VF detection in dual-chamber ICDs. In a single chamber detection scenario for a wider population, as in the case of prophylactic ICDs, it is expected that rapidly conducted AF that conducts at ventricular rates that overlap with the VF zone will also be a primary algorithmic cause for inappropriate detection. See, Wilkoff B. L. et. al., “Critical Analysis of Dual-Chamber Implantable Cardioverter-Defibrillator Arrhythmia Detection: Results and Technical Consideration”, Circulation, 2001; 103:381-386.
The QRS morphology during rapidly conducted AF often differs from the QRS morphology during NSR rendering algorithms that rely on a finding of similarity between the current QRS complex morphology and an NSR QRS complex morphology to distinguish SVT from VF much less effective. Although the QRS morphology during AF episodes differs from NSR QRS morphology, there is often a characteristic QRS complex morphology during AF that is relatively stable over short periods of time.
Therefore, a need remains for a robust and computationally efficient VF detection capability of discriminating a true VF episode from high rate VT or SVT that is not life threatening particularly for use in prophylactic ICDs to avoid the unnecessary delivery of a C/D shock therapy. Such a VF detection capability would, of course, still be beneficial in more complex single chamber, dual chamber and multi-chamber ICDs. Such a robust VF detection capability may also find utility in an implantable heart monitors (IHM) having a sense electrode array (SEA) implanted subcutaneously for monitoring, processing, and storing data from the EGM sensed across one or more selected far field sense vector as described in commonly assigned U.S. Pat. No. 5,331,966, for example.
Moreover, a need remains for a robust and computationally efficient AF detection capability of discriminating a true AF episode from high rate AT.